#!/bin/bash
# vLLM server for one config (runs in its own tmux session). Usage: tb_server.sh CONFIG GPU PORT
# Other models (defaults = Qwen3.8-27B): KVQ_MODEL KVQ_TB_NAME KVQ_REASONING_PARSER, KVQ_SERVE_ARGS adds serve args.
CFG=$1; GPU=$2; PORT=$3
case $CFG in
  BF16)    KV=bfloat16; BE=FLASH_ATTN ;;
  BF16-FI) KV=bfloat16; BE=FLASHINFER ;;
  FP8)     KV=fp8_e4m3; BE=FLASH_ATTN ;;
  FP4)     KV=nvfp4;    BE=FLASHINFER ;;
esac
source ~/kvq-env/bin/activate
eval "SERVE_ARGS=(${KVQ_SERVE_ARGS:-})"
mkdir -p ~/tbench/servers
MODEL=${KVQ_MODEL:-~/models/Qwen3.8-27B}; NAME=${KVQ_TB_NAME:-qwen3.8-27b}; RP=${KVQ_REASONING_PARSER:-qwen3}
echo "== vLLM $CFG kv=$KV backend=$BE gpu=$GPU port=$PORT $(date -u)"
CUDA_VISIBLE_DEVICES=$GPU exec vllm serve $MODEL --served-model-name $NAME --port $PORT \
  --dtype bfloat16 --max-model-len 262144 --gpu-memory-utilization 0.90 --reasoning-parser $RP --seed 0 \
  --kv-cache-dtype $KV --attention-backend $BE "${SERVE_ARGS[@]}" 2>&1 | tee -a ~/tbench/servers/${KVQ_TB_LOG_PREFIX:-}$CFG.log